Mention rate
The proportion of eligible completed responses that recognisably mention the business. A mention need not be an affirmative recommendation.
London Auto Index / Research division
EDITION 1 · IN VALIDATIONAI visibility.
Measured with context.
The UK Automotive AI Visibility Index™ is being developed to examine how automotive businesses appear in a defined sample of AI-assisted buyer questions. It measures answer behaviour within a research window—not dealership quality.
No final Edition 1 rankings are published. Any figures previously shown have been withdrawn pending validation. The preview below demonstrates reporting with fictional data.
Defined prompts and research windowMentions, recommendations and citationsNo pay-for-placement or score changes
Publication status
The method is published.
The rankings are not final.
Methodology draft version 0.9 sets out the proposed research design. Precise scores and rankings depend on reproducible prompt runs, platform records, repeat tests and quality checks.
- Methodology
PUBLISHED DRAFTVersion 0.9, published 15 September 2026. Proposed weights and rules remain subject to validation. - Edition 1
IN VALIDATIONThe evidence archive must support the prompt runs, research dates, platform coverage, repeat results and reviewed classifications. - Public rankings
NOT FINALNo precise Edition 1 rankings are published here. A future edition must disclose sample sizes, missing data and known limitations. - Report preview
FICTIONAL DATAA demonstration of what a report can contain. It is not a live AI test, client outcome or Index scorecard.
Proposed measurement framework
Four components.
One bounded research question.
How visible is a business within the eligible buyer questions during a stated research window? These are the proposed composite-score components from methodology v0.9.
Recommendation rank
A normalised position score for placement within an AI-generated recommendation set. It is distinct from a simple recommendation rate.
Citation share
The proportion of eligible completed responses with a supporting link to the business’s verified domain. A citation does not by itself establish a recommendation.
Consistency
The stability of presence across repeat runs of the same prompt-platform combination. Repeats help expose answer variation.
Provisional formula: 0.40M + 0.30R + 0.20C + 0.10S. Weights, normalisation and evidence requirements must be versioned before final rankings are published. The fictional preview uses example reporting measures and does not calculate this composite score.
Explore the format
See the evidence.
Understand the uncertainty.
Change the example business type, then choose a buyer question. Every business, answer and figure in this demonstration is fictional.
Interactive dealer report preview
What would your AI visibility assessment show?
Choose a business type to see the evidence, uncertainty and next action a dealer report can contain.
Illustrative example only. All businesses, answers and figures here are fictional. This preview does not run a live AI test or report an Index ranking. Edition 1 remains in validation.
Example design: 20 buyer prompts × 2 platforms × 2 repeats; 76 of 80 planned answers completed. This is a demonstration of reporting, not a claimed research run.
Mention rateCorrect name in answers
Recommendation rateAffirmative recommendations
Owned citation rateAnswers linking to the dealer site
Repeat agreementPairs agreeing on mention status
Example prompt evidence
Select a buyer question to inspect the example observation and next check.
Diagnosis and delivery
Relevant peer view:
First action:
1 · Freeze promptsBuyer intent, London scope and peer definition
2 · Capture and reviewAnswers, recommendations, citations and entity facts
3 · RetestSame prompts and settings; disclose missing answers and variation
A real report carries collection dates, exact denominators, archived evidence, coverage, volatility and fairness context. A high AI visibility rate is not a dealer quality rating. Use the assessment form below to discuss a separately scoped measurement.
Edition 1 research scope
Platform coverage
must be disclosed.
The proposed scope covers six products. This list describes intended coverage, not confirmation that all planned tests have been completed.
ChatGPTGeminiClaudePerplexityGrokGoogle AI Overviews
A published edition must identify platforms successfully tested, visible product or model versions where available, collection dates and access limitations.
What makes a run reviewable?
- Stable prompt IDs, verbatim questions and eligibility rules.
- Target design of three repeats for each eligible prompt-platform combination.
- Archived answers, timestamps, platform context and citation URLs.
- Refusals, errors and incomplete outputs recorded separately.
- Reviewed entity matches and totals reconciled to run-level records.
Reading the Index responsibly
Visibility is an observation.
It is not an endorsement.
The Index cannot establish the internal reason an AI system produced an answer. It observes outputs under documented conditions.
Does a higher score mean a better dealership?
No. AI visibility does not establish service quality, vehicle condition, trustworthiness or commercial superiority. A benchmark is a time-bounded observation of answer behaviour.
Can a business pay for a higher score?
No. Commissioning a diagnostic or advisory service cannot purchase inclusion, placement or a score change in the public Index.
How will missing answers be handled?
Errors, refusals and incomplete outputs are recorded separately. A published edition must disclose its minimum completed sample. Businesses below that threshold should be marked insufficient data rather than given a precise rank.
Does the preview test my dealership?
No. Changing the example type only changes pre-written fictional data. It does not query an AI platform, inspect your business or generate a live score.
Can I challenge an identification error?
Businesses can submit relevant evidence of identification or classification errors. Material corrections should be dated; commercial payment cannot determine the outcome.
Separately scoped commercial assessment
What would you like
to understand?
A private assessment can examine your dealership’s mentions, citations, entity accuracy and buyer-question gaps under an agreed scope. It is separate from the public Index validation process.
Share your location, vehicle specialisms and priorities. The next step is a conversation about fit, platforms and deliverables.
No guaranteed AI recommendation, ranking or sale. Commissioning an assessment does not buy public Index placement. Your enquiry is not a purchase or research-result claim.
LAI / Assessment enquiry
No automatic marketing subscription. Please provide business information relevant to your request.
Published LAI research
Explore the questions
behind the measurement.
Citations and brand mentions
Why a linked source and a named business need to be recorded separately.
Read the measurement protocol →London buyer consideration
A founding brief on the buyer journey and a proposed measurement approach.
Read the buyer brief →The research hub
Browse published work and distinguish it from research still in development.
Explore LAI research →